- Email: [email protected]
Experimental study on impact behavior of submarine landslides on undersea communication cables
Ocean Engineering 148 (2018) 530–537
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Experimental study on impact behavior of submarine landslides on undersea communication cables Fawu Wang, Zili Dai *, Yasutaka Nakahara, Tomokazu Sonoyama Department of Geoscience, Shimane University, Matsue, Shimane 690-8504, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: Submarine landslide Undersea cable Impact force Physical model test Turbidity current
Submarine landslides, which are characterized by large scale and long run-out distance, could destroy undersea communication cables, thus resulting in a great number of economic losses. In this work, an experimental apparatus is designed to simulate the relative motion between the submarine landslide and undersea communication cable, and investigate the impact behavior of a submarine landslide. The test results show that submarine landslide with larger thickness and larger particle size can generate larger impact force on the communication cables. With an increase of the sliding velocity, the evolution of the flow behavior and impact force of the soil-water mixture can be divided into three stages: 1) landslide stage, the mixture consists a water layer and a sand layer, the impact force increases with the sliding velocity and reaches a peak value at this stage; 2) transforming stage, sand particles start to be eroded by water, and result into a turbidity current layer above the soil-water interface. At this stage, the impact force decreases to a minimum value; 3) turbidity current stage, all of the sand particles are eroded by water, and the impact force increases again when increasing the velocity continuously.
1. Introduction Submarine communication cables are laid on the sea bed between land-based stations to carry digital data across stretches of ocean with speed and security, including telephone, Internet and private data traffic. The total length of submarine communication cables in the world's oceans is over 1 million km (Carter et al., 2009). These cables are usually buried 1 m and exceptionally up to 10 m beneath the seafloor, or placed on the seabed where water depths are greater than 1500 m. The undersea environment is unfavorable for the submarine communication cables. For example, continental shelves are typically exposed to wave and current actions that move seabed sediment and result in the exposure or even undermining of a submarine cable. Submarine geological activities, such as earthquakes, mud volcanoes, submarine landslides and turbidity currents, pose serious natural threat to the submarine communication cables (Kvalstad et al., 2001). Submarine landslides are perhaps the most significant of them, which occur frequently on continental margins and slopes, releasing huge sediment volumes that may travel distances as long as hundreds of kilometers. They could easily damage submarine cables and lead to interruption of data transmission and international communications. According to the statistics (Carter et al., 2009), there were 2162 cable
breaks globally between 1960 and 2006, of which at least 20% were directly influenced by submarine landslides or turbidity currents. One serious event occurred on 19 November 1929. A strong earthquake shake the continental slopes south of the Grand Banks, off the eastern coast of the United States and Canada, and caused a lot of submarine landslides on the continental slopes. Some of these slope failures transformed into high-velocity turbidity currents, which traveled over 720 km from the source area. In 13 h following the earthquake, submarine communication cables near the earthquake epicenter were broken in sequence from north to south with the passage of time (Heezen and Ewing, 1952; Hasegawa and Kanamori, 1987; Canals et al., 2004; Fine et al., 2005). Since then, similar cases have been reported around the world, especially in earthquake-prone regions. For example, Krause et al. (1970) recorded an earthquake-triggered submarine landslide which disrupted a telephone cable at least 280 km away in water depths over 6600 m in the New Britain Trench in 1966. The 2003 Boumerdes earthquake in Algeria damaged six cables and disrupted all submarine communication networks in the Mediterranean region (Joseph and Hussong, 2003). More recently, as a consequence of submarine landslides and turbidity currents associated with the 2006 Pingtung earthquakes offshore Southwest Taiwan, eleven submarine cables across the Kaoping canyon and Manila trench were broken in
* Corresponding author. E-mail address: [email protected] (Z. Dai). https://doi.org/10.1016/j.oceaneng.2017.11.050 Received 4 June 2017; Received in revised form 16 November 2017; Accepted 26 November 2017 0029-8018/© 2017 Published by Elsevier Ltd.
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
rotation, three types of sensors are fixed inside the frame bed at a line just below the cable and paralleling to the axle, so that the three sensors can reach the lowest point at the same time with the cable: 1) Load cell used to measure the weight of water and soil when the sensor comes to the lowest point, then normal stress can be calculated; 2) Shear stress sensor used to measure the shear resistance at the bottom of the soil and water mixture; 3) Water pressure transducer used to measure the water pressure when the sensor reaches the lowest point. The sensors are distributed in section A just below the cable, as shown in Fig. 1(b). Rotation velocity of the trough is controlled by a motor. A data logger and battery are fixed in the axial part of the apparatus, see Fig. 1(b), so that data collection is carried out in an independent unit during rotation.
sequence from 1500 to 4000 m depth (Hsu et al., 2008). The interruptions in international telecommunication affected all the East and Southeast Asian countries. The violent cable failures that happened in these cases are evidence of the destructive power of the submarine landslides. Therefore, the propagation of submarine landslides is one of the most serious threats to undersea communication cables. Investigation on the impact force of submarine landslides during the propagation is of great importance to maintain the safety of submarine communication cables. Due to the extreme undersea environment, it is nearly impossible to directly observe submarine landslides and measure the impact force, therefore, the investigation on post-failure behavior of submarine landslides is a big challenge. Most of the previous studies used the flume tests. For example, Zakeri et al. (2008) designed and conducted a series of flume tests to investigate the impact force exerted by a clay-rich submarine landslide on two pipelines. Based on the experimental results, they proposed a method to estimate the drag forces normal to the pipeline axis. Combining with numerical analysis, Zakeri (2009) proposed a method to estimate the normal and longitudinal drag forces on a suspended pipeline. Based on Zakeri's work, Randolph and White (2012) proposed a failure envelope to estimate the interaction force between offshore pipelines and submarine landslides. However, the available literature is still insufficient in providing a comprehensive method to estimate the impact forces of submarine landslide and turbidity current on seafloor installations, such as communication cables and oil/gas pipelines. In the presented work, an annular trough which can rotate in a vertical plane at a controlled velocity is developed and used to simulate the relative motion between a submarine landslide and an undersea cable. The main advantage of this apparatus is that the sliding velocity of the soil-water mixture could be easily controlled by adjusting the rotation velocity of the trench. Through the tests, the flow behaviour of the soil-water mixture at different velocities (from a submarine landslide to a turbidity current) is observed, and its impact force exerting on an underwater cable is investigated. The influence of the landslide volume, soil type and sliding velocity is considered and discussed.
2.2. Materials used in the tests In this study, fine silica sands no. 7 (S7) and no. 8 (S8) were used to simulate the submarine landslide mass. Their physical properties are presented in Table 1. Fresh water was used in this study. 2.3. Test procedure In this study, four types of tests were carried out to investigate the impact force of submarine landslides, and the effect of landslide sliding velocity, landslide volume, soil types, and diameter of the cable are considered and discussed. The details are listed below: 1) Tests were conducted at 13 different constant velocities (the linear velocity of the trough bottom), i.e., 0.02, 0.07, 0.13, 0.20, 0.26, 0.32, 0.38, 0.45, 0.51, 0.58, 0.65, 0.72, and 0.77 m/s. Three circles were rotated at each velocity, because the stable impacting can be achieved at this state. Noting that the maximum linear velocity of the trough was about 0.77 m/s due to the output power limitation of the motor. Therefore, the velocity range of the tests was set to be 0.02–0.77 m/s, though the maximum velocity recorded for real submarine landslides could be much larger (Heezen et al., 1954). 2) Tests were conducted using four different constant masses and keeping the water level at the same height. The dry mass quantities are 10, 20, 30, and 40 kg, respectively. 3) To investigate the effect of soil characteristic, two types of soils (silica sand no.7 and no.8) were used in the tests, and the results were compared. 4) Cable with three different diameters of 0.01, 0.02 and 0.03 m were used to evaluate the effect of cable dimension. The material of the cables were the same stainless.
2. Experimental program 2.1. Experimental setup Fig. 1 shows the experimental apparatus to simulate a submarine landslide and the relative motion between the submarine landslide and undersea cable. The main part of the apparatus is an annular trough, with an axle in the centre. The outer radius, inner radius and width of the trough are 0.9 m, 0.6 m and 0.4 m, respectively. Its frame is made of steel and the front is transparent plexiglas, through which the motion of the sand and water inside the trough could be easily observed. The axle of the apparatus is connected to a motor so that it could rotate in a vertical plane at a controlled velocity. The sketch of the test apparatus is shown in Fig. 1(b). A certain amount of water and soil material are placed in the apparatus to simulate the submarine landslide. In this paper, the term “submarine landslide” is used to refer to the subaqueous mass motion; the soil-water interface is obvious in this status. With the increase of the flowing velocity, the water and soil material are mixed completely and the interface disappears, this flowing mixture is defined as “turbidity current”. A cable is installed in the bottom of the annular trough paralleling to the axel. The cable is made of stainless steel with a length of 370 mm. Two load cells are set at two sides of the cable to measure the impact force from the water and landslide mass, as shown in Fig. 1(b). The height of the cable from the frame bottom could be easily adjusted. During the rotation of the trough, the soil-water mixture always stay in the lower part of the apparatus under the action of gravity, and the relative motion between the apparatus bottom and cable shows in a similar way to a real submarine landslide on the seabed. To monitor the behavior of the soil and water mixture during
In these tests, the bottom of the cable was set at a height of 0.02 m above the trough bottom (point A in Fig. 1). The maximum depth of the water was kept at 0.2 m for all tests. For comparison purpose, water-onlytests and sand-only-tests were conducted at first. 3. Test results In this work, the cable is installed paralleling to the axel of the annular trough and perpendicular to the flow direction of the submarine landslide. Therefore, the “impact force” discussed in this section refers to the normal force exerting on the cables due to the submarine landslide propagation. The tangential force along the axel of the cable is not discussed in this work. 3.1. Water-only-tests Water-only-tests were conducted at first, and the impact forces on the cable due to water were investigated at different rotation velocities. In these tests, the diameter of the cable was 0.02 m. Fig. 2 shows the results 531
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 1. Test apparatus to simulate a submarine landslide; (a) Picture of the test apparatus, (b) Sketch showing water, soil and cable in the apparatus and sensor setting.
which nearly appeared at the same time with the peak values of water pressure and normal stress. Fig. 3 shows the relationship between the impact force with the rotation velocity. With the increase of the rotation velocity, the impact force increased, and the increasing rate became larger.
of water-only tests (rotation velocity was 0.38 m/s). The blue and green lines represent the water pressure and normal stress respectively. The peak values were both about 1.98 kPa, generated by 0.20 mm water head, thus verifying the precision of the sensors. The red line was the impact force F acting on the cable. The peak value of F was about 3.16 N,
3.2. Sand-only -tests
Table 1 Properties of the silica sand no. 7 and no. 8. Properties
S7
S8
Specific gravity, Gs Mean grain size, D50 (mm) Effective grain size, D10 (mm)
2.640 0.130 0.074
2.640 0.050 0.018
For the purpose of comparison, sand-only-tests were conducted to investigate the impact force from dry sand flow on the cable. Fig. 4 shows the test result with 30 kg silica sand no. 7. The diameter of the cable was 20 mm, and the rotation velocity was 0.38 m/s. In this figure, the red curve represented the impact force, and its maximum value was about 532
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 2. Test results with only water (D ¼ 20 mm, v ¼ 0.38 m/s).
Fig. 5. Impact force due to dry sand at different rotation velocity (D ¼ 20 mm).
velocity reached a certain value, the sand mass became very loose, and behaved just like a fluid along the bottom of the trough. This phenomenon was defined as phase transform in the literatures (Esipov and €schel, 1997; Huang et al., 2015). When increasing the sand mass from Po 10 kg to 40 kg, the impact force obviously increased. Therefore, the impact force of the dry sand flow depended on the sand mass and flowing velocity to a significant extent. 3.3. Submarine landslide tests To simulate submarine landslides, sand and water were mixed in the trough. The flow behaviour of the soil-water mixture is shown in Fig. 6. The state of soil-water mixture during the landslide propagation can be divided into three stages with an increase of the sliding velocity: 1) landslide stage, the mixture consists a water layer and a sand layer, the interface between the sand and water was clear, see Fig. 6(a); 2) transforming stage, sand particles start to be eroded by the ambient water, and a turbidity current layer appears above the sand-water interface, as shown in Fig. 6(b); 3) turbidity current stage, all of the sand particles are eroded by water and the sand layer disappears, see Fig. 6(c). Fig. 7 shows the test result with 30 kg silica sand no. 7 at the velocity of 0.38 m/s. The cable diameter in this case was 20 mm. As shown in the figure, the peak value of the impact force, normal stress and water pressure were 22.4 N, 2.53 kPa, and 2.04 kPa, respectively. Comparing this result with Fig. 4, the impact force from the sand-water mixture is much smaller than dry sand. Therefore, when the landslide volume and the soil property are the same, the impact force of the subaqueous landslide is smaller than the subaerial landslide. Fig. 8 shows the impact force from the sand-water mixture at different velocities. The diameters of the cables were 10, 20, and 30 mm, respectively. The sand used in the figures was no. 7 silica sand. At the same velocity, the impact force increased with the sand mass (from 10 to 40 kg), as shown in the three figures. When increasing the sand volume, the thickness of the sand layer in the trench increased, thus resulting in the large impact force. Note that the landslide thickness may vary substantially in the trench, therefore, the impact force discussed in this paper refers to the peak value during the three circles. As shown in Fig. 8, when increasing the velocity, the impact force increased at first, and then reached a peak value. For example, in Fig. 8(a), when the no. 7 silica sand was 10, 20, 30, and 40 kg, the cable with diameter of 10 mm was subjected to a maximum impact force of 9.13, 12.02, 15.46, and 19.47 N respectively from the sand-water mixture. Then, the impact force experienced a significant decrease and reached a minimum value. The minimum impact forces on the cable were
Fig. 3. Impact force due to water at different rotation velocity.
Fig. 4. Test results with dry sand (S7, m ¼ 30 kg, D ¼ 20 mm, v ¼ 0.38 m/s).
30.2 N. The green curve was the normal stress generated by dry sand. The peak value was around 1.76 kPa. Pore water pressure was not measured in sand-only tests. Fig. 5 shows the impact forces of the dry sand flow under different velocities. With the rotation velocity increasing, the impact force increased at first stage, and than followed by a slight decrease. When the 533
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 6. Flow behaviour of the submarine landslide at different velocities; (a) v ¼ 0.02 m/s, (b) v ¼ 0.32 m/s, (c) v ¼ 0.65 m/s.
difference between the normal stress and water pressure, and could be calculated by:
3.32, 4.97, 6.49, and 7.87 N when the no. 7 silica sand was 10, 20, 30 and 40 kg. After that, the impact force increased again with the velocity. It is worth noting that because of the limitation of the motor power, the range of the rotation velocity of the trough is 0.02–0.77 m/s. The maximum and minimum impact forces discussed in this paper referred to the values in this velocity range. Fig. 9 shows the evolution of the effective stress during the test. In this work, the effective stress of the sand was defined as the
σ0 ¼ σn u
(1)
where σ n is the normal stress, u is the pore water pressure. Noting that the normal stress and pore water pressure were measured by the sensor installed at the bottom center of the trough and under the cable, as
534
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 7. Test results with sand-water mixture (S7, m ¼ 30 kg, D ¼ 20 mm, v ¼ 0.38 m/s).
shown in Fig. 1(b). With the increase of the sand mass (from 10 to 40 kg), the effective stress significantly increased, therefore the impact force increased obviously. On the other hand, with increasing the rotation velocity, the sand particle gradually eroded by the ambient water, the effective stress decreased, leading to the decrease of the impact force. 4. Discussion To investigate the impact force per unit area on the cables, the concept of distributed load DL is introduced, which can be calculated by:
DL ¼ F=A
(2)
where F is the impact force and A is the cross sectional area of the cable. There the distributed load DL has the dimension of pressure kN/m2. For the cables with diameter of 10, 20, and 30 mm, the cross-sectional area A equals 0.0037, 0.0074, and 0.0111 m2 respectively. The relationship between the distributed load DL with the rotation velocity is shown in Fig. 10. The three sets of data match with each other basically. Therefore, the dimension of the cable only affected the impact force, but made almost no difference to distributed load. As shown in Fig. 8, the evolution of the impact force with the velocity can be divided into three stages. At the first stage, the impact force increased and reached a peak value, and the velocity was defined as critical velocity VI, about 0.2 m/s. At this stage, the sand mass moved slowly on the trough bottom as a landslide mass, the interface between the sand and water was quite easy to be identified, as shown in Fig. 6(a). Then, the impact force decreased to a minimum value, the velocity at this moment was defined as the critical velocity VII, about 0.58 m/s. At this stage, sand particles were gradually eroded by the water, the water became turbid, and the water-sand interface became obscure, as shown in Fig. 6(b). At the last stage, the impact force increased again when increasing the sliding velocity. All of the sand particles were mixed with the water and resulted in a turbidity current, see Fig. 6(c). Therefore, it can be concluded that the impact force of the submarine landslide is related to the state of the soil-water mixture. In the third stage, the sand-water mixture becomes a turbidity current. At that time, the distributed load depends largely on the suspended-sediment concentrations of the turbidity current, which is defined as the ratio of the sand quantity to the volume of the turbidity current. Supposing that the sand particles were evenly distributed in the water, the suspended-sediment concentrations of the turbidity current
Fig. 8. Impact force of the sand-water mixture under different velocity; (a) D ¼ 10 mm, (b) D ¼ 20 mm, (c) D ¼ 30 mm.
535
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 11. Distributed load of the sand-water mixture with different suspended-sediment concentrations.
Fig. 9. Evolution of the effective stress with the rotation velocity.
White (2012) proposed a formula to estimate the normal loading of a steady state flow acting on a pipeline, which showed a positive correlation relation between the impact force and flowing velocity. Therefore, based on the results of this work and those reported in the literatures, the impact force of a submarine landslide was in proportion to the square of the velocity when the sand-water mixture resulted into a turbidity current. Fig. 13 compares the distributed load on the cable (D ¼ 20 mm) from the mixture with silica sand no. 7 and no. 8. When the sand mass quantity was 10, 20, 30, and 40 kg, the maximum impact force from the silica sand no. 8 was 1.54, 2.26, 2.75 and 3.05 kN/mm2, respectively. While for the submarine landslide with silica sand no. 7, the maximum impact force was 1.90, 2.71, 3.35 and 4.01 kN/mm2. Therefore, the submarine landslide with larger grain size can generate larger impact force. This figure also shows that the critical velocities of no. 8 silica sand were a little smaller than no. 7 silica sand. For no. 8 silica sand, the distributed load reached maximum value at about 0.07 m/s, and the minimum value at about 0.45 m/s. For no. 7 silica sand, the critical velocity VI was about 0.20 m/s and VII was about 0.60 m/s. As shown in the above results, this work investigated the impact force of the submarine landslide acting on a cable through a series of tests. The influence from the landslide volume, sliding velocity, soil type of the landslide, and cable diameter was discussed. In this work, the cable
Fig. 10. Distributed load on the cables with different diameter.
were 0.16, 0.32, 0.48, and 0.64 kg/l for the tests with 10, 20, 30, and 40 kg silica sand. For the water-only tests, the suspended-sediment concentration is 0 kg/l. Fig. 11 shows the distributed load on the cable exerted by the turbidity currents with different suspended-sediment concentrations. In general, the distributed load is proportional to the concentration of the turbidity current, as shown by the curve fitting in Fig. 11. Similar with the impact behavior in water-only tests, the impact force of a turbidity current increased with the velocity. Fig. 12 shows the relationship between the distributed load with the rotation velocity at the third stage. 10 kg silica sand no. 8 was used in this case, and the cable diameter was 20 mm. The distributed load was nearly proportional to the square of the velocity. Noting that when the rotation velocity was smaller than 0.38 m/s, sand particles started to precipitate and the suspendedsediment concentration of the turbidity current changed. Therefore, the distributed loads with the velocity smaller than 0.38 m/s were not included in Fig. 12. In the existing literatures, turbidity current was usually considered to be a Non-Newtonian fluid (Zakeri et al., 2008; Zakeri, 2009; Randolph and White, 2012). Impact behavior of a non-Newtonian fluid flow was firstly investigated by Pazwash and Robertson (1975) based on fluid dynamics and rheology principles. They stated that the impact force of a Non-Newtonian fluid was in direct proportion to the square of the velocity. More recently, Randolph and
Fig. 12. Distributed load on the cable under different rotation velocity. 536
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
3) The evolution of the impact force with the velocity could be divided into three stages. It increased and reached a peak value at the first stage, and then decreased to a minimum value at the second stage. At the last stage, the impact force increased again when increasing the velocity continuously. 4) The submarine landslide resulted into a turbidity current at the third stage. At this moment, the impact force was in proportion to the suspended-sediment concentration and the square of the velocity; 5) The impact force from the sand-water mixture with the coarse sands was larger than that with the finer ones. Therefore, submarine landslides with larger particle size could act larger impact force on the cables. Acknowledgements This work was supported by the Japanese Scientific Research Grant (No. 20310109). Fig. 13. Distributed load from the sand with different particle sizes.
References Carter, L., Burnett, D., Drew, S., Marle, G., Hagadorn, L., Bartlett-McNeil, D., Irvine, N., 2009. Submarine Cables and the Oceans: Connecting the World. UNEP-WCMC Biodiversity Series No. 31. ICPC/UNEP/UNEP-WCMC. Canals, M., Lastras, G., Urgeles, R., Casamor, J.L., Mienert, J., Cattaneo, A., De Batist, M., Haflidason, H., Imbo, Y., Laberg, J.S., Locat, J., Long, D., Longva, O., Masson, D.G., Sultan, N., Trincardi, F., Bryn, P., 2004. Slope failure dynamics and impacts from seafloor and shallow sub-seafloor geophysical data: case studies from the COSTA project. Mar. Geol. 213 (1–4), 9–72. Esipov, S.E., P€ oschel, T., 1997. The granular phase diagram. J. Stat. Phys. 86 (5–6), 1385–1395. Fine, I.V., Rabinovich, A.B., Bornhold, B.D., Thomson, R.E., Kulikov, E.A., 2005. The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Mar. Geol. 215 (1), 45–57. Hasegawa, H.S., Kanamori, H., 1987. Source mechanism of the magnitude 7.2 Grand Banks earthquake of November 1929: double couple or submarine landslide? Bull. Seismol. Soc. Am. 77 (6), 1984–2004. Heezen, B.C., Ewing, M., 1952. Turbidity currents and submarine slumps, and the 1929 Grand Banks earthquake. Am. J. Sci. 250, 849–873. Heezen, B.C., Ericson, D.B., Ewing, M., 1954. Further evidence for a turbidity current following the 1929 Grand Banks earthquake. Deep Sea Res. (1953) 1 (4), 193–202. Hsu, S.K., Kuo, J., Lo, C.L., Tsai, C.H., Doo, W.B., Ku, C.Y., Sibuet, J.C., 2008. Turbidity currents, submarine landslides and the 2006 Pingtung Earthquake off SWTaiwan. Terr. Atmos. Ocean Sci. 19 (6), 762–772. Huang, Y., Zhu, C., Xiang, X., Mao, W., 2015. Liquid-gas-like phase transition in sand flow under microgravity. Microgravity Sci. Technol. 27 (3), 155–170. Joseph, D., Hussong, D.M., 2003. Geospatial management of commercial seafloor data. In: 2003 ESRI User Conference Proceedings. Krause, D.C., White, W.C., Piper, D.J.W., Heezen, B.C., 1970. Turbidity currents and cable breaks in the western New Britain Trench. Geol. Soc. Am. Bull. 81, 2153–2160. Kvalstad, T.J., Nadim, F., Arbitz, C.B., 2001. Deepwater geohazards: geotechnical concerns and solutions. In: Offshore Technology Conference. Pazwash, H., Robertson, J.M., 1975. Forces on bodies in bingham fluids. J. Hydraul. Res. 13 (1), 35–55. Randolph, M.F., White, D.J., 2012. Interaction forces between pipelines and submarine slides—a geotechnical viewpoint. Ocean Eng. 48, 32–37. Zakeri, A., Høeg, K., Nadim, F., 2008. Submarine debris flow impact on pipelines—Part I: experimental investigation. Coast. Eng. 55 (12), 1209–1218. Zakeri, A., 2009. Submarine debris flow impact on suspended (free-span) pipelines: normal and longitudinal drag forces. Ocean Eng. 36 (6), 489–499.
paralleled to the axel of the trough. The impact angle was always perpendicular to flow direction of the submarine landslide. Therefore, the impact force discussed in this paper mainly referred to the normal force. The tangential force along the axel of the cable was not included in this work. An important consideration in model test is the presence of an artificial boundary in the test system due to the containment walls. In the presented work, the empty trough space between the outer and inner rings is very small, therefore, the walls may influence the flow behaviour of the soil-water mixture during the tests. However, the presence of the water in the trench can decrease the friction between the sand and walls during the tests, therefore can reduce the boundary effect. 5. Conclusions This work presented the preliminary results of the experimental study on impact force of submarine landslides. Since the dimension of the test apparatus was minute compared to real submarine landslides, it was quite difficult to obtain quantitative results through these tests. However, some promising qualitative tendencies related to the impact force on cables from submarine landslides were presented, which could be concluded as follows: 1) Cables with larger diameter subjected larger impact force from the submarine landslides, but the distributed load on the cable was similar; 2) Effective stress generated by the landslide mass was one of the key factor to influence the impact behavior. Submarine landslide with larger thickness could generate larger effective stress, and then lead to larger impact force on the cables;
537
Contents lists available at ScienceDirect
Ocean Engineering journal homepage: www.elsevier.com/locate/oceaneng
Experimental study on impact behavior of submarine landslides on undersea communication cables Fawu Wang, Zili Dai *, Yasutaka Nakahara, Tomokazu Sonoyama Department of Geoscience, Shimane University, Matsue, Shimane 690-8504, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: Submarine landslide Undersea cable Impact force Physical model test Turbidity current
Submarine landslides, which are characterized by large scale and long run-out distance, could destroy undersea communication cables, thus resulting in a great number of economic losses. In this work, an experimental apparatus is designed to simulate the relative motion between the submarine landslide and undersea communication cable, and investigate the impact behavior of a submarine landslide. The test results show that submarine landslide with larger thickness and larger particle size can generate larger impact force on the communication cables. With an increase of the sliding velocity, the evolution of the flow behavior and impact force of the soil-water mixture can be divided into three stages: 1) landslide stage, the mixture consists a water layer and a sand layer, the impact force increases with the sliding velocity and reaches a peak value at this stage; 2) transforming stage, sand particles start to be eroded by water, and result into a turbidity current layer above the soil-water interface. At this stage, the impact force decreases to a minimum value; 3) turbidity current stage, all of the sand particles are eroded by water, and the impact force increases again when increasing the velocity continuously.
1. Introduction Submarine communication cables are laid on the sea bed between land-based stations to carry digital data across stretches of ocean with speed and security, including telephone, Internet and private data traffic. The total length of submarine communication cables in the world's oceans is over 1 million km (Carter et al., 2009). These cables are usually buried 1 m and exceptionally up to 10 m beneath the seafloor, or placed on the seabed where water depths are greater than 1500 m. The undersea environment is unfavorable for the submarine communication cables. For example, continental shelves are typically exposed to wave and current actions that move seabed sediment and result in the exposure or even undermining of a submarine cable. Submarine geological activities, such as earthquakes, mud volcanoes, submarine landslides and turbidity currents, pose serious natural threat to the submarine communication cables (Kvalstad et al., 2001). Submarine landslides are perhaps the most significant of them, which occur frequently on continental margins and slopes, releasing huge sediment volumes that may travel distances as long as hundreds of kilometers. They could easily damage submarine cables and lead to interruption of data transmission and international communications. According to the statistics (Carter et al., 2009), there were 2162 cable
breaks globally between 1960 and 2006, of which at least 20% were directly influenced by submarine landslides or turbidity currents. One serious event occurred on 19 November 1929. A strong earthquake shake the continental slopes south of the Grand Banks, off the eastern coast of the United States and Canada, and caused a lot of submarine landslides on the continental slopes. Some of these slope failures transformed into high-velocity turbidity currents, which traveled over 720 km from the source area. In 13 h following the earthquake, submarine communication cables near the earthquake epicenter were broken in sequence from north to south with the passage of time (Heezen and Ewing, 1952; Hasegawa and Kanamori, 1987; Canals et al., 2004; Fine et al., 2005). Since then, similar cases have been reported around the world, especially in earthquake-prone regions. For example, Krause et al. (1970) recorded an earthquake-triggered submarine landslide which disrupted a telephone cable at least 280 km away in water depths over 6600 m in the New Britain Trench in 1966. The 2003 Boumerdes earthquake in Algeria damaged six cables and disrupted all submarine communication networks in the Mediterranean region (Joseph and Hussong, 2003). More recently, as a consequence of submarine landslides and turbidity currents associated with the 2006 Pingtung earthquakes offshore Southwest Taiwan, eleven submarine cables across the Kaoping canyon and Manila trench were broken in
* Corresponding author. E-mail address: [email protected] (Z. Dai). https://doi.org/10.1016/j.oceaneng.2017.11.050 Received 4 June 2017; Received in revised form 16 November 2017; Accepted 26 November 2017 0029-8018/© 2017 Published by Elsevier Ltd.
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
rotation, three types of sensors are fixed inside the frame bed at a line just below the cable and paralleling to the axle, so that the three sensors can reach the lowest point at the same time with the cable: 1) Load cell used to measure the weight of water and soil when the sensor comes to the lowest point, then normal stress can be calculated; 2) Shear stress sensor used to measure the shear resistance at the bottom of the soil and water mixture; 3) Water pressure transducer used to measure the water pressure when the sensor reaches the lowest point. The sensors are distributed in section A just below the cable, as shown in Fig. 1(b). Rotation velocity of the trough is controlled by a motor. A data logger and battery are fixed in the axial part of the apparatus, see Fig. 1(b), so that data collection is carried out in an independent unit during rotation.
sequence from 1500 to 4000 m depth (Hsu et al., 2008). The interruptions in international telecommunication affected all the East and Southeast Asian countries. The violent cable failures that happened in these cases are evidence of the destructive power of the submarine landslides. Therefore, the propagation of submarine landslides is one of the most serious threats to undersea communication cables. Investigation on the impact force of submarine landslides during the propagation is of great importance to maintain the safety of submarine communication cables. Due to the extreme undersea environment, it is nearly impossible to directly observe submarine landslides and measure the impact force, therefore, the investigation on post-failure behavior of submarine landslides is a big challenge. Most of the previous studies used the flume tests. For example, Zakeri et al. (2008) designed and conducted a series of flume tests to investigate the impact force exerted by a clay-rich submarine landslide on two pipelines. Based on the experimental results, they proposed a method to estimate the drag forces normal to the pipeline axis. Combining with numerical analysis, Zakeri (2009) proposed a method to estimate the normal and longitudinal drag forces on a suspended pipeline. Based on Zakeri's work, Randolph and White (2012) proposed a failure envelope to estimate the interaction force between offshore pipelines and submarine landslides. However, the available literature is still insufficient in providing a comprehensive method to estimate the impact forces of submarine landslide and turbidity current on seafloor installations, such as communication cables and oil/gas pipelines. In the presented work, an annular trough which can rotate in a vertical plane at a controlled velocity is developed and used to simulate the relative motion between a submarine landslide and an undersea cable. The main advantage of this apparatus is that the sliding velocity of the soil-water mixture could be easily controlled by adjusting the rotation velocity of the trench. Through the tests, the flow behaviour of the soil-water mixture at different velocities (from a submarine landslide to a turbidity current) is observed, and its impact force exerting on an underwater cable is investigated. The influence of the landslide volume, soil type and sliding velocity is considered and discussed.
2.2. Materials used in the tests In this study, fine silica sands no. 7 (S7) and no. 8 (S8) were used to simulate the submarine landslide mass. Their physical properties are presented in Table 1. Fresh water was used in this study. 2.3. Test procedure In this study, four types of tests were carried out to investigate the impact force of submarine landslides, and the effect of landslide sliding velocity, landslide volume, soil types, and diameter of the cable are considered and discussed. The details are listed below: 1) Tests were conducted at 13 different constant velocities (the linear velocity of the trough bottom), i.e., 0.02, 0.07, 0.13, 0.20, 0.26, 0.32, 0.38, 0.45, 0.51, 0.58, 0.65, 0.72, and 0.77 m/s. Three circles were rotated at each velocity, because the stable impacting can be achieved at this state. Noting that the maximum linear velocity of the trough was about 0.77 m/s due to the output power limitation of the motor. Therefore, the velocity range of the tests was set to be 0.02–0.77 m/s, though the maximum velocity recorded for real submarine landslides could be much larger (Heezen et al., 1954). 2) Tests were conducted using four different constant masses and keeping the water level at the same height. The dry mass quantities are 10, 20, 30, and 40 kg, respectively. 3) To investigate the effect of soil characteristic, two types of soils (silica sand no.7 and no.8) were used in the tests, and the results were compared. 4) Cable with three different diameters of 0.01, 0.02 and 0.03 m were used to evaluate the effect of cable dimension. The material of the cables were the same stainless.
2. Experimental program 2.1. Experimental setup Fig. 1 shows the experimental apparatus to simulate a submarine landslide and the relative motion between the submarine landslide and undersea cable. The main part of the apparatus is an annular trough, with an axle in the centre. The outer radius, inner radius and width of the trough are 0.9 m, 0.6 m and 0.4 m, respectively. Its frame is made of steel and the front is transparent plexiglas, through which the motion of the sand and water inside the trough could be easily observed. The axle of the apparatus is connected to a motor so that it could rotate in a vertical plane at a controlled velocity. The sketch of the test apparatus is shown in Fig. 1(b). A certain amount of water and soil material are placed in the apparatus to simulate the submarine landslide. In this paper, the term “submarine landslide” is used to refer to the subaqueous mass motion; the soil-water interface is obvious in this status. With the increase of the flowing velocity, the water and soil material are mixed completely and the interface disappears, this flowing mixture is defined as “turbidity current”. A cable is installed in the bottom of the annular trough paralleling to the axel. The cable is made of stainless steel with a length of 370 mm. Two load cells are set at two sides of the cable to measure the impact force from the water and landslide mass, as shown in Fig. 1(b). The height of the cable from the frame bottom could be easily adjusted. During the rotation of the trough, the soil-water mixture always stay in the lower part of the apparatus under the action of gravity, and the relative motion between the apparatus bottom and cable shows in a similar way to a real submarine landslide on the seabed. To monitor the behavior of the soil and water mixture during
In these tests, the bottom of the cable was set at a height of 0.02 m above the trough bottom (point A in Fig. 1). The maximum depth of the water was kept at 0.2 m for all tests. For comparison purpose, water-onlytests and sand-only-tests were conducted at first. 3. Test results In this work, the cable is installed paralleling to the axel of the annular trough and perpendicular to the flow direction of the submarine landslide. Therefore, the “impact force” discussed in this section refers to the normal force exerting on the cables due to the submarine landslide propagation. The tangential force along the axel of the cable is not discussed in this work. 3.1. Water-only-tests Water-only-tests were conducted at first, and the impact forces on the cable due to water were investigated at different rotation velocities. In these tests, the diameter of the cable was 0.02 m. Fig. 2 shows the results 531
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 1. Test apparatus to simulate a submarine landslide; (a) Picture of the test apparatus, (b) Sketch showing water, soil and cable in the apparatus and sensor setting.
which nearly appeared at the same time with the peak values of water pressure and normal stress. Fig. 3 shows the relationship between the impact force with the rotation velocity. With the increase of the rotation velocity, the impact force increased, and the increasing rate became larger.
of water-only tests (rotation velocity was 0.38 m/s). The blue and green lines represent the water pressure and normal stress respectively. The peak values were both about 1.98 kPa, generated by 0.20 mm water head, thus verifying the precision of the sensors. The red line was the impact force F acting on the cable. The peak value of F was about 3.16 N,
3.2. Sand-only -tests
Table 1 Properties of the silica sand no. 7 and no. 8. Properties
S7
S8
Specific gravity, Gs Mean grain size, D50 (mm) Effective grain size, D10 (mm)
2.640 0.130 0.074
2.640 0.050 0.018
For the purpose of comparison, sand-only-tests were conducted to investigate the impact force from dry sand flow on the cable. Fig. 4 shows the test result with 30 kg silica sand no. 7. The diameter of the cable was 20 mm, and the rotation velocity was 0.38 m/s. In this figure, the red curve represented the impact force, and its maximum value was about 532
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 2. Test results with only water (D ¼ 20 mm, v ¼ 0.38 m/s).
Fig. 5. Impact force due to dry sand at different rotation velocity (D ¼ 20 mm).
velocity reached a certain value, the sand mass became very loose, and behaved just like a fluid along the bottom of the trough. This phenomenon was defined as phase transform in the literatures (Esipov and €schel, 1997; Huang et al., 2015). When increasing the sand mass from Po 10 kg to 40 kg, the impact force obviously increased. Therefore, the impact force of the dry sand flow depended on the sand mass and flowing velocity to a significant extent. 3.3. Submarine landslide tests To simulate submarine landslides, sand and water were mixed in the trough. The flow behaviour of the soil-water mixture is shown in Fig. 6. The state of soil-water mixture during the landslide propagation can be divided into three stages with an increase of the sliding velocity: 1) landslide stage, the mixture consists a water layer and a sand layer, the interface between the sand and water was clear, see Fig. 6(a); 2) transforming stage, sand particles start to be eroded by the ambient water, and a turbidity current layer appears above the sand-water interface, as shown in Fig. 6(b); 3) turbidity current stage, all of the sand particles are eroded by water and the sand layer disappears, see Fig. 6(c). Fig. 7 shows the test result with 30 kg silica sand no. 7 at the velocity of 0.38 m/s. The cable diameter in this case was 20 mm. As shown in the figure, the peak value of the impact force, normal stress and water pressure were 22.4 N, 2.53 kPa, and 2.04 kPa, respectively. Comparing this result with Fig. 4, the impact force from the sand-water mixture is much smaller than dry sand. Therefore, when the landslide volume and the soil property are the same, the impact force of the subaqueous landslide is smaller than the subaerial landslide. Fig. 8 shows the impact force from the sand-water mixture at different velocities. The diameters of the cables were 10, 20, and 30 mm, respectively. The sand used in the figures was no. 7 silica sand. At the same velocity, the impact force increased with the sand mass (from 10 to 40 kg), as shown in the three figures. When increasing the sand volume, the thickness of the sand layer in the trench increased, thus resulting in the large impact force. Note that the landslide thickness may vary substantially in the trench, therefore, the impact force discussed in this paper refers to the peak value during the three circles. As shown in Fig. 8, when increasing the velocity, the impact force increased at first, and then reached a peak value. For example, in Fig. 8(a), when the no. 7 silica sand was 10, 20, 30, and 40 kg, the cable with diameter of 10 mm was subjected to a maximum impact force of 9.13, 12.02, 15.46, and 19.47 N respectively from the sand-water mixture. Then, the impact force experienced a significant decrease and reached a minimum value. The minimum impact forces on the cable were
Fig. 3. Impact force due to water at different rotation velocity.
Fig. 4. Test results with dry sand (S7, m ¼ 30 kg, D ¼ 20 mm, v ¼ 0.38 m/s).
30.2 N. The green curve was the normal stress generated by dry sand. The peak value was around 1.76 kPa. Pore water pressure was not measured in sand-only tests. Fig. 5 shows the impact forces of the dry sand flow under different velocities. With the rotation velocity increasing, the impact force increased at first stage, and than followed by a slight decrease. When the 533
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 6. Flow behaviour of the submarine landslide at different velocities; (a) v ¼ 0.02 m/s, (b) v ¼ 0.32 m/s, (c) v ¼ 0.65 m/s.
difference between the normal stress and water pressure, and could be calculated by:
3.32, 4.97, 6.49, and 7.87 N when the no. 7 silica sand was 10, 20, 30 and 40 kg. After that, the impact force increased again with the velocity. It is worth noting that because of the limitation of the motor power, the range of the rotation velocity of the trough is 0.02–0.77 m/s. The maximum and minimum impact forces discussed in this paper referred to the values in this velocity range. Fig. 9 shows the evolution of the effective stress during the test. In this work, the effective stress of the sand was defined as the
σ0 ¼ σn u
(1)
where σ n is the normal stress, u is the pore water pressure. Noting that the normal stress and pore water pressure were measured by the sensor installed at the bottom center of the trough and under the cable, as
534
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 7. Test results with sand-water mixture (S7, m ¼ 30 kg, D ¼ 20 mm, v ¼ 0.38 m/s).
shown in Fig. 1(b). With the increase of the sand mass (from 10 to 40 kg), the effective stress significantly increased, therefore the impact force increased obviously. On the other hand, with increasing the rotation velocity, the sand particle gradually eroded by the ambient water, the effective stress decreased, leading to the decrease of the impact force. 4. Discussion To investigate the impact force per unit area on the cables, the concept of distributed load DL is introduced, which can be calculated by:
DL ¼ F=A
(2)
where F is the impact force and A is the cross sectional area of the cable. There the distributed load DL has the dimension of pressure kN/m2. For the cables with diameter of 10, 20, and 30 mm, the cross-sectional area A equals 0.0037, 0.0074, and 0.0111 m2 respectively. The relationship between the distributed load DL with the rotation velocity is shown in Fig. 10. The three sets of data match with each other basically. Therefore, the dimension of the cable only affected the impact force, but made almost no difference to distributed load. As shown in Fig. 8, the evolution of the impact force with the velocity can be divided into three stages. At the first stage, the impact force increased and reached a peak value, and the velocity was defined as critical velocity VI, about 0.2 m/s. At this stage, the sand mass moved slowly on the trough bottom as a landslide mass, the interface between the sand and water was quite easy to be identified, as shown in Fig. 6(a). Then, the impact force decreased to a minimum value, the velocity at this moment was defined as the critical velocity VII, about 0.58 m/s. At this stage, sand particles were gradually eroded by the water, the water became turbid, and the water-sand interface became obscure, as shown in Fig. 6(b). At the last stage, the impact force increased again when increasing the sliding velocity. All of the sand particles were mixed with the water and resulted in a turbidity current, see Fig. 6(c). Therefore, it can be concluded that the impact force of the submarine landslide is related to the state of the soil-water mixture. In the third stage, the sand-water mixture becomes a turbidity current. At that time, the distributed load depends largely on the suspended-sediment concentrations of the turbidity current, which is defined as the ratio of the sand quantity to the volume of the turbidity current. Supposing that the sand particles were evenly distributed in the water, the suspended-sediment concentrations of the turbidity current
Fig. 8. Impact force of the sand-water mixture under different velocity; (a) D ¼ 10 mm, (b) D ¼ 20 mm, (c) D ¼ 30 mm.
535
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
Fig. 11. Distributed load of the sand-water mixture with different suspended-sediment concentrations.
Fig. 9. Evolution of the effective stress with the rotation velocity.
White (2012) proposed a formula to estimate the normal loading of a steady state flow acting on a pipeline, which showed a positive correlation relation between the impact force and flowing velocity. Therefore, based on the results of this work and those reported in the literatures, the impact force of a submarine landslide was in proportion to the square of the velocity when the sand-water mixture resulted into a turbidity current. Fig. 13 compares the distributed load on the cable (D ¼ 20 mm) from the mixture with silica sand no. 7 and no. 8. When the sand mass quantity was 10, 20, 30, and 40 kg, the maximum impact force from the silica sand no. 8 was 1.54, 2.26, 2.75 and 3.05 kN/mm2, respectively. While for the submarine landslide with silica sand no. 7, the maximum impact force was 1.90, 2.71, 3.35 and 4.01 kN/mm2. Therefore, the submarine landslide with larger grain size can generate larger impact force. This figure also shows that the critical velocities of no. 8 silica sand were a little smaller than no. 7 silica sand. For no. 8 silica sand, the distributed load reached maximum value at about 0.07 m/s, and the minimum value at about 0.45 m/s. For no. 7 silica sand, the critical velocity VI was about 0.20 m/s and VII was about 0.60 m/s. As shown in the above results, this work investigated the impact force of the submarine landslide acting on a cable through a series of tests. The influence from the landslide volume, sliding velocity, soil type of the landslide, and cable diameter was discussed. In this work, the cable
Fig. 10. Distributed load on the cables with different diameter.
were 0.16, 0.32, 0.48, and 0.64 kg/l for the tests with 10, 20, 30, and 40 kg silica sand. For the water-only tests, the suspended-sediment concentration is 0 kg/l. Fig. 11 shows the distributed load on the cable exerted by the turbidity currents with different suspended-sediment concentrations. In general, the distributed load is proportional to the concentration of the turbidity current, as shown by the curve fitting in Fig. 11. Similar with the impact behavior in water-only tests, the impact force of a turbidity current increased with the velocity. Fig. 12 shows the relationship between the distributed load with the rotation velocity at the third stage. 10 kg silica sand no. 8 was used in this case, and the cable diameter was 20 mm. The distributed load was nearly proportional to the square of the velocity. Noting that when the rotation velocity was smaller than 0.38 m/s, sand particles started to precipitate and the suspendedsediment concentration of the turbidity current changed. Therefore, the distributed loads with the velocity smaller than 0.38 m/s were not included in Fig. 12. In the existing literatures, turbidity current was usually considered to be a Non-Newtonian fluid (Zakeri et al., 2008; Zakeri, 2009; Randolph and White, 2012). Impact behavior of a non-Newtonian fluid flow was firstly investigated by Pazwash and Robertson (1975) based on fluid dynamics and rheology principles. They stated that the impact force of a Non-Newtonian fluid was in direct proportion to the square of the velocity. More recently, Randolph and
Fig. 12. Distributed load on the cable under different rotation velocity. 536
F. Wang et al.
Ocean Engineering 148 (2018) 530–537
3) The evolution of the impact force with the velocity could be divided into three stages. It increased and reached a peak value at the first stage, and then decreased to a minimum value at the second stage. At the last stage, the impact force increased again when increasing the velocity continuously. 4) The submarine landslide resulted into a turbidity current at the third stage. At this moment, the impact force was in proportion to the suspended-sediment concentration and the square of the velocity; 5) The impact force from the sand-water mixture with the coarse sands was larger than that with the finer ones. Therefore, submarine landslides with larger particle size could act larger impact force on the cables. Acknowledgements This work was supported by the Japanese Scientific Research Grant (No. 20310109). Fig. 13. Distributed load from the sand with different particle sizes.
References Carter, L., Burnett, D., Drew, S., Marle, G., Hagadorn, L., Bartlett-McNeil, D., Irvine, N., 2009. Submarine Cables and the Oceans: Connecting the World. UNEP-WCMC Biodiversity Series No. 31. ICPC/UNEP/UNEP-WCMC. Canals, M., Lastras, G., Urgeles, R., Casamor, J.L., Mienert, J., Cattaneo, A., De Batist, M., Haflidason, H., Imbo, Y., Laberg, J.S., Locat, J., Long, D., Longva, O., Masson, D.G., Sultan, N., Trincardi, F., Bryn, P., 2004. Slope failure dynamics and impacts from seafloor and shallow sub-seafloor geophysical data: case studies from the COSTA project. Mar. Geol. 213 (1–4), 9–72. Esipov, S.E., P€ oschel, T., 1997. The granular phase diagram. J. Stat. Phys. 86 (5–6), 1385–1395. Fine, I.V., Rabinovich, A.B., Bornhold, B.D., Thomson, R.E., Kulikov, E.A., 2005. The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Mar. Geol. 215 (1), 45–57. Hasegawa, H.S., Kanamori, H., 1987. Source mechanism of the magnitude 7.2 Grand Banks earthquake of November 1929: double couple or submarine landslide? Bull. Seismol. Soc. Am. 77 (6), 1984–2004. Heezen, B.C., Ewing, M., 1952. Turbidity currents and submarine slumps, and the 1929 Grand Banks earthquake. Am. J. Sci. 250, 849–873. Heezen, B.C., Ericson, D.B., Ewing, M., 1954. Further evidence for a turbidity current following the 1929 Grand Banks earthquake. Deep Sea Res. (1953) 1 (4), 193–202. Hsu, S.K., Kuo, J., Lo, C.L., Tsai, C.H., Doo, W.B., Ku, C.Y., Sibuet, J.C., 2008. Turbidity currents, submarine landslides and the 2006 Pingtung Earthquake off SWTaiwan. Terr. Atmos. Ocean Sci. 19 (6), 762–772. Huang, Y., Zhu, C., Xiang, X., Mao, W., 2015. Liquid-gas-like phase transition in sand flow under microgravity. Microgravity Sci. Technol. 27 (3), 155–170. Joseph, D., Hussong, D.M., 2003. Geospatial management of commercial seafloor data. In: 2003 ESRI User Conference Proceedings. Krause, D.C., White, W.C., Piper, D.J.W., Heezen, B.C., 1970. Turbidity currents and cable breaks in the western New Britain Trench. Geol. Soc. Am. Bull. 81, 2153–2160. Kvalstad, T.J., Nadim, F., Arbitz, C.B., 2001. Deepwater geohazards: geotechnical concerns and solutions. In: Offshore Technology Conference. Pazwash, H., Robertson, J.M., 1975. Forces on bodies in bingham fluids. J. Hydraul. Res. 13 (1), 35–55. Randolph, M.F., White, D.J., 2012. Interaction forces between pipelines and submarine slides—a geotechnical viewpoint. Ocean Eng. 48, 32–37. Zakeri, A., Høeg, K., Nadim, F., 2008. Submarine debris flow impact on pipelines—Part I: experimental investigation. Coast. Eng. 55 (12), 1209–1218. Zakeri, A., 2009. Submarine debris flow impact on suspended (free-span) pipelines: normal and longitudinal drag forces. Ocean Eng. 36 (6), 489–499.
paralleled to the axel of the trough. The impact angle was always perpendicular to flow direction of the submarine landslide. Therefore, the impact force discussed in this paper mainly referred to the normal force. The tangential force along the axel of the cable was not included in this work. An important consideration in model test is the presence of an artificial boundary in the test system due to the containment walls. In the presented work, the empty trough space between the outer and inner rings is very small, therefore, the walls may influence the flow behaviour of the soil-water mixture during the tests. However, the presence of the water in the trench can decrease the friction between the sand and walls during the tests, therefore can reduce the boundary effect. 5. Conclusions This work presented the preliminary results of the experimental study on impact force of submarine landslides. Since the dimension of the test apparatus was minute compared to real submarine landslides, it was quite difficult to obtain quantitative results through these tests. However, some promising qualitative tendencies related to the impact force on cables from submarine landslides were presented, which could be concluded as follows: 1) Cables with larger diameter subjected larger impact force from the submarine landslides, but the distributed load on the cable was similar; 2) Effective stress generated by the landslide mass was one of the key factor to influence the impact behavior. Submarine landslide with larger thickness could generate larger effective stress, and then lead to larger impact force on the cables;
537